PROJECT SUMMARY/ABSTRACT Persons with opioid use disorder (OUD) have longer and more expensive hospitalizations (over $15 billion annually) than general hospital admissions. Despite the magnitude of this issue, little research describes hospital use of evidence-based opioid agonist treatments (OATs). I propose to assess hospital variation, model associations with patient and hospital attributes, and explore hospital structures influencing OAT delivery in the context of local and broader environments. There are three OAT (continuation of pre-hospital OAT, OAT for withdrawal management, and OAT initiation for long-term care) and three non-OAT (short-acting opioids, symptom management medications, and other care) OUD care scenarios. I will measure patient disposition and 30-day readmissions for each care scenario. This study leverages health service delivery conceptual frameworks and a multilevel change model using a mixed methods approach. The study addresses two quantitative aims (analysis of OUD hospital admissions in the Premier Healthcare Database) and one qualitative aim (interviews with 20 hospital key informants from a purposive sample of 10 to 12 hospitals). Specific Aim 1. Assess variation in hospital OAT delivery and test the associations of OAT and non- OAT care scenarios with attributes and outcomes. I will evaluate whether meaningful variation exists in hospital OAT delivery and whether patient attributes (e.g., gender), hospital attributes (e.g., size) and patient outcomes (e.g., disposition) are associated with the provision of OAT in the hospital. H1a: Measurable differences exist among hospital OAT quartiles and between top and bottom OAT quartiles. H1b: Hospital OAT quartiles will be associated with at least one patient attribute, one hospital attribute, and one patient outcome domain. Specific Aim 2. Model and test the proposed relationships of patient and hospital attributes on hospital OAT delivery and patient outcomes. I will use multilevel modeling to identify specific associations of patient (e.g., race) and hospital (e.g., teaching status) attributes with hospital OAT variation, patient outcomes, and between-hospital variation. H2a: OAT delivery will be negatively associated with patient race (non-white), insurance status (uninsured), and hospital teaching status (non-teaching). H2b: In-hospital mortality and 30-day readmissions will be negatively associated with hospital OAT delivery. Specific Aim 3. Use qualitative interviews with key hospital informants to explore the environmental context and hospital structures that may mediate hospital OAT delivery. Findings from Aims 1 and 2 guide the purposive sampling and interview protocol for the key informant interviews. I will probe for local hospital (e.g., addiction medicine fellowship) and broader environmental contextual factors (e.g., federal policy) that may influence hospital structures (e.g., formulary) impeding or facilitating hospital OAT delivery.